U.S. patent application number 17/648710 was filed with the patent office on 2022-07-07 for testing switches in a power converter.
The applicant listed for this patent is pSemi Corporation. Invention is credited to Gregory SZCZESZYNSKI, Brian ZANCHI.
Application Number | 20220214393 17/648710 |
Document ID | / |
Family ID | |
Filed Date | 2022-07-07 |
United States Patent
Application |
20220214393 |
Kind Code |
A1 |
SZCZESZYNSKI; Gregory ; et
al. |
July 7, 2022 |
TESTING SWITCHES IN A POWER CONVERTER
Abstract
A switching network includes a switch, a driver for the switch,
and a floating-regulator that powers the driver. The
floating-regulator includes a shunt that is used only when testing
the network. The shunt diverts biasing current so that it does not
interfere with a measurement of an electrical property of a
switch.
Inventors: |
SZCZESZYNSKI; Gregory;
(Hollis, NH) ; ZANCHI; Brian; (Dracut,
MA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
pSemi Corporation |
San Diego |
CA |
US |
|
|
Appl. No.: |
17/648710 |
Filed: |
January 24, 2022 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
16023163 |
Jun 29, 2018 |
11262395 |
|
|
17648710 |
|
|
|
|
International
Class: |
G01R 31/26 20060101
G01R031/26; H02M 3/07 20060101 H02M003/07; G01R 31/40 20060101
G01R031/40; H02M 1/08 20060101 H02M001/08 |
Claims
1. An apparatus comprising a switching network that, when connected
to capacitors, forms a switched-capacitor network for transforming
a first voltage into a second voltage, wherein said switching
network comprises a switch, a driver, and a floating-regulator,
wherein said floating-regulator comprises a first current path and
a second current path, wherein current that proceeds along said
first current path drives said switch, wherein current that
proceeds along said second path is diverted such that said current
fails to provide power for driving any switch, and wherein said
floating-regulator comprises a shunt for causing said current to
proceed along one of said first and second paths.
2. The apparatus of claim 1, further comprising a controller that
is configured for causing said shunt to select said second
path.
3. The apparatus of claim 1, wherein, during operation, said
switching network comprises a node of lower electrical potential,
wherein said node of lower electrical potential has an electrical
potential that is lower than all other electric potentials in said
switching network, and wherein said second path leads to said node
of lowest electrical potential.
4. The apparatus of claim 1, further comprising a controller
configured to determine whether a switch in said switching network
has a leakage current that complies with a design specification,
wherein said controller is configured to cause a bias current in
said switch, to divert said bias current so that said bias current
avoids interfering with measurement of said leakage current, to
apply a voltage across said switch, to obtain a measurement of a
current that flows in response to application of said voltage, to
derive said leakage current from said measurement, to determine
that said leakage current falls outside of said design
specification, and to identify said switching network as being
defective.
5. The apparatus of claim 1, further comprising a power converter,
wherein said switching network and said capacitors define a
switched-capacitor network that is a constituent of said power
converter, and wherein said power converter comprises a controller
for controlling said switching network.
6. The apparatus of claim 1, further comprising a power converter,
wherein said switching network and said capacitors define a
switched-capacitor network that is a constituent of said power
converter, and wherein, during operation of said power converter,
said shunt never switches between said first and second paths.
7. The apparatus of claim 1, wherein said switching network
comprises a first set of switches and a second set of switches,
wherein switches from said first set of switches are arranged in
series to define a first charge-transfer path that extends from an
input of said switching network to an output of said switching
network, wherein switches from said second set of switches are
arranged in series to define a second charge-transfer path that
extends from said input to said output, wherein said first
charge-transfer path comprises nodes at which switches from said
first set of switches connect to each other, wherein said nodes are
configured to be connected to a first plurality of capacitors,
wherein said second charge-transfer path comprises nodes at which
switches from said second set of switches connect to each other,
and wherein said nodes are configured to be connected to a second
plurality of capacitors.
8. The apparatus of claim 1, further comprising a monolithic
substrate, wherein said switching circuit is integrated into said
monolithic substrate.
9. The apparatus of claim 1, further comprising first and second
monolithic substrates, wherein said switching circuit is integrated
into said first monolithic substrate and said capacitors are
integrated into said second monolithic substrate, and wherein said
first and second monolithic substrates are interconnected.
10. The apparatus of claim 1, further comprising a monolithic
substrate and a plurality of capacitors, wherein said switching
circuit is integrated into said monolithic substrate, and wherein
said capacitors are connected to said switching circuit.
11. The apparatus of claim 1, wherein said switch comprises a power
FET.
12. The apparatus of claim 1, wherein said switch is one of a
plurality of switches that are distributed between first and second
charge-transfer paths that both extend between an input and an
output of said switching network, and wherein said
floating-regulator is cross-coupled between said first and second
charge-transfer paths.
13. The apparatus of claim 1, wherein said floating-regulator
comprises a regulator switch to regulate flow of charge into said
driver and a Zener diode, wherein said Zener diode causes said flow
of charge to be provided at a fixed offset from a floating
voltage.
14. The apparatus of claim 1, wherein said floating-regulator
comprises a regulator switch to regulate flow of charge into said
driver and a plurality of conducting paths for supplying currents
to said regulator switch, wherein said conducting paths extend
between different sources of charge and said regulator switch.
15. The apparatus of claim 1, wherein said floating-regulator
comprises a first diode, a second diode, and a regulator switch,
wherein said first diode connects said regulator switch to a first
source of charge and wherein said second diode connects said
regulator switch to a second source of charge.
16. The apparatus of claim 1, wherein at least some of current that
proceeds along said first current path and that drives said switch
is provided to an additional floating-regulator to provide power
for driving an additional switch.
17. The apparatus of claim 1, further comprising a controller
configured to estimate a leakage current of a switch in said
switching network, said controller being configured to cause a bias
current to bias said switch to be in a non-conducting state and to
divert said bias current so that said bias current avoids
interfering with measurement of said leakage current.
18. A method comprising testing a leakage current of a switch in a
switching network in an integrated circuit, that, when connected to
capacitors, forms a switched-capacitor network for transforming a
first voltage into a second voltage, wherein said switching network
comprises a driver and a floating-regulator, wherein testing said
leakage current comprises causing a first current, at least some of
which will be used to bias said switch, diverting said first
current along a path such that said first current avoids
interference with a measurement of a second current, obtaining a
measurement of said second current, and estimating said leakage
current based at least in part based on said measurement of said
second current.
19. The method of claim 18, wherein said switch is one of a
plurality of switches that are in series along a charge-transfer
path, and wherein diverting said first current comprises causing
said first current to avoid flowing along said charge-transfer
path.
20. The method of claim 18, further comprising connecting an
external voltage source across said switch, wherein said external
voltage source is external to said switched-capacitor network,
wherein said second current is current that enters said switching
network.
21. The method of claim 18, further comprising suppressing the
ability of said second current from flowing along a path other than
a charge-transfer path that leads to said switch.
22. The method of claim 18, wherein said switch is a first switch
that is in series with a second switch along a charge-transfer
path, wherein said method further comprises suppressing said
current that enters said switching network from flowing along said
charge-transfer path towards said second switch.
23. The method of claim 18, wherein said first current flows
through said floating-regulator, wherein said floating-regulator
controls driving of said switch.
24. The method of claim 18, wherein said first current carries
power, and wherein diverting said first current comprises wasting
said power.
25-33. (canceled)
Description
FIELD OF INVENTION
[0001] This invention relates to testing of an integrated circuit,
and in particular, to testing switches for a switched-capacitor
circuit.
BACKGROUND
[0002] A power converter often includes a switched-capacitor
circuit that converts one voltage into another voltage. The
switched-capacitor circuit generally includes a switching network
having one or more switches that are used to interconnect various
capacitors. When these switches are closed, they handle significant
amounts of current.
[0003] The switches are typically implemented as field-effect
transistors. A voltage that is applied to a gate terminal of the
transistor closes this switch. This voltage controls the existence
of a conducting channel between the transistor's source and drain
terminals. Like any conducting channel, this channel will have some
resistance to current flow. Because this channel is between the
transistor's drain and source, and because it is established in the
transistor's "ON" state, this resistance is often called the
"RDSON" of the transistor.
[0004] The value of RDSON is of great significance. This is
particularly so when the switch is called upon to handle
significant amounts of current. It is therefore important that the
channel's resistance conform to the design specification.
[0005] Another parameter of interest is the transistor's leakage
current. Even when no conducting channel connects the source and
drain across the gate, it is still possible for a few charge
carriers to diffuse from the source all the way to the drain. This
results in a small leakage current.
[0006] Both the transistor's RDSON and its leakage current are
subject to variability as a result of imperfections in the
manufacturing process. Thus, it is important to measure both RDSON
and leakage current as a part of the manufacturing process. Doing
so provides the opportunity to reject defective switches and to
thereby avoid the likelihood of placing a faulty power converter
into the stream of commerce.
SUMMARY
[0007] In one aspect, the invention features an apparatus that
includes a switching network. The switching network includes
switches, drivers for those switches, and floating-regulators that
power the drivers. Each floating-regulator includes a shunt that is
used only when testing the network. The shunt diverts biasing
current so that it does not interfere with a measurement of an
electrical property of a switch.
[0008] In another aspect, the invention features a switching
network that, when connected to capacitors, forms a
switched-capacitor network for transforming a first voltage into a
second voltage. The switching network includes a plurality of
switches, a corresponding plurality of drivers, and a corresponding
plurality of floating-regulators. Among these are first and second
switches, a first driver, and a first floating-regulator that has
first and second current paths, together with a shunt for selecting
between them. Current that proceeds along the first current path is
provided to a second floating-regulator to provide power for
driving the second switch. However, current that proceeds along the
second path is diverted. As a result, this current fails to provide
power for driving any switch.
[0009] Some embodiments include a controller that causes the shunt
to select the second path.
[0010] In other embodiments, during operation, the switching
network comprises a node of lower electrical potential. This node
has an electrical potential that is lower than all other electric
potentials in the switching network. In such embodiments, the
second path leads to this node of lower electrical potential. In
other embodiments, this node of lower electrical potential is the
node of lowest electrical potential.
[0011] Yet other embodiments include a controller configured to
determine whether a switch in the switching network has an
electrical parameter that complies with a design specification.
Such a controller causes a bias current in the switch and diverts
the bias current so that it avoids interfering with measurement of
the electrical parameter. Such a controller also applies an
electrical stimulus, measures an electrical response, derives the
electrical parameter from the electrical response, and determines
that the electrical parameter falls outside of the design
specification and identifies the switching network as being
defective.
[0012] Also among the embodiments are those that include a
controller configured to measure an extent to which a switch
conducts electric current when the switch is held in a conducting
state. Such a controller causes a bias current to hold the switch
in the conducting state, diverts the bias current so that the bias
current avoids interfering with a known current that is injected
through the switch, causes injection of the known current through
the switch, measures a voltage across the switch, and determines
the extent to which the switch conducts based on the measurement of
the voltage and the known current.
[0013] Yet other embodiments include a controller configured to
determine whether a switch in the switching network has an
electrical parameter that complies with a design specification. In
such embodiments, the controller causes a bias current in the
switch, diverts the bias current so that the bias current avoids
interfering with measurement of the electrical parameter, measures
the electrical parameter, determines that the electrical parameter
falls outside of the design specification, and identifies the
switching network as being defective.
[0014] Still other embodiments include a controller configured to
determine a leakage current of a switch in the switching network.
Such a controller causes a bias current to bias the switch to be in
a non-conducting state and diverts the bias current so that the
bias current avoids interfering with measurement of the leakage
current.
[0015] Other embodiments include a power converter. In these
embodiments, the switching network and capacitors define a
switched-capacitor network that is a constituent element of the
power converter. Among these embodiments are those that also
include a controller that controls the switches. Also among these
embodiments are those in which the shunt never switches between the
first and second paths.
[0016] In other embodiments, the plurality of switches comprises a
first set of switches that are arranged in series to define a first
charge-transfer path that extends from the switching network's
input to its output. This first charge-transfer path comprises
nodes at which switches from the first set of switches connect.
These nodes are also configured to be connected to the
capacitors.
[0017] In other embodiments, the plurality of switches comprises
first and second sets of switches. Switches from the first set are
arranged in series to define a first charge-transfer path and
switches from the second set are likewise arranged in series to
define a second charge-transfer path. Both charge-transfer paths
extend from the switching network's input to its output. The first
charge-transfer path comprises nodes at which switches from the
first set of switches connect to each other. These nodes are
configured to be connected to a first plurality of capacitors.
Meanwhile, the second charge-transfer path comprises nodes at which
switches from the second set of switches connect to each other.
These nodes are configured to be connected to a second plurality of
capacitors.
[0018] Some embodiments also include a monolithic substrate into
which the switching circuit is integrated.
[0019] Alternative embodiments have first and second monolithic
substrates. In these embodiments, the switching circuit is
integrated into the first monolithic substrate and the capacitors
are integrated into the second monolithic substrate. The first and
second monolithic substrates are interconnected so that the
switching circuit is electrically connected to the capacitors.
[0020] Some embodiments include a monolithic substrate and a
plurality of capacitors. In these embodiments, the switching
circuit is integrated into the monolithic substrate and the
capacitors are connected to the switching circuit.
[0021] Other embodiments include those in which the switches
comprise power FETs.
[0022] In yet other embodiments, each switch has multiple parallel
paths between its drain and its source, with each such path being
individually controlled.
[0023] Additional embodiments include those in which the switches
are distributed between first and second charge-transfer paths that
both extend between an input and an output of the switching
network. In these embodiments, the floating-regulators are
cross-coupled between the first and second charge-transfer
paths.
[0024] Still other embodiments are those in which the first
floating-regulator comprises a regulator switch to regulate flow of
charge into the first driver and a Zener diode that causes the flow
of charge to be provided at a fixed offset from a floating
voltage.
[0025] Additional embodiments include those in which the
floating-regulator comprises a regulator switch to regulate flow of
charge into the first driver and conducting paths for supplying
currents to the regulator switch. These conducting paths extend
between different sources of charge and the regulator switch.
[0026] Also among the embodiments are those in which the first
floating-regulator comprises a regulator switch and diodes that
connect the regulator switch to corresponding sources of electric
charge.
[0027] In another embodiment, the first floating-regulator includes
a regulator switch to regulate flow of charge into the first driver
and a Zener diode. The Zener diode causes the flow of charge to be
provided at a fixed offset from a floating voltage.
[0028] Also among the embodiments are those in which the first
floating-regulator includes a regulator switch to regulate flow of
charge into the first driver and a plurality of conducting paths
for supplying currents to the regulator switch. These the
conducting paths extend between different sources of charge and the
regulator switch. Examples of different sources of charge include
different capacitors to which the switching network is connected
and also an external power supply that is outside the
switched-capacitor network altogether.
[0029] In some embodiments, the first floating-regulator includes a
regulator switch and first and second diodes. The diodes connect
the regulator switch to different sources of charge.
[0030] In yet other embodiments, the first floating-regulator
includes a device that permits to current flow unless the potential
imposed across its terminals is of a first polarity and has a
magnitude less than a threshold. Such embodiments feature a
regulator switch that regulates current into a first driver and
also into the device.
[0031] In another aspect, the invention features a method
comprising testing an electrical parameter of a switch in a
switching network in an integrated circuit, that, when connected to
capacitors, forms a switched-capacitor network for transforming a
first voltage into a second voltage, the switching network having a
plurality of switches, a corresponding plurality of drivers, and a
corresponding plurality of floating-regulators. In such a method,
testing the electrical parameter comprises causing a first current,
at least some of which will be used to bias the switch, diverting
the current along a path such that the current avoids interference
with a measurement, and carrying out the measurement.
[0032] In some practices, the method includes, after having
diverted the current, measuring a leakage current of the
switch.
[0033] In other practices, the method includes, measuring an RDSON
of the switch.
[0034] Other practices further include providing a stimulus to the
switch, measuring a response to the stimulus, and based on the
response, determining the electrical parameter.
[0035] In another aspect, the invention includes testing a leakage
current of a first switch in a switching network in an integrated
circuit, that, when connected to capacitors, forms a
switched-capacitor network for transforming a first voltage into a
second voltage. The switching network includes a plurality of
switches, one of which is the first switch. It also includes a
corresponding plurality of drivers, and a corresponding plurality
of floating-regulators. Testing the leakage current includes
causing a first current, at least some of which will be used to
bias the first switch, diverting it along a path such that it
avoids interference with a measurement of a second current,
measuring the second current, and determining the leakage current
at least in part based on the measurement of the second
current.
[0036] In some practices, the method comprises connecting an
external voltage source across the switch. Such an external voltage
source is external to the switched-capacitor network. As such, it
excludes the capacitors of the switched-capacitor network itself.
The second current is current that enters the switching
network.
[0037] Additional practices include suppressing the ability of the
second current from flowing along a path other than a
charge-transfer path that leads to the first switch.
[0038] In those practices in which the first switch is in series
with a second switch on the charge-transfer path, the method also
includes suppressing the current that enters the switching network
from flowing along the charge-transfer path towards the second
switch.
[0039] In yet other practices, one of the floating-regulators
controls driving of the first switch. Among these practices are
those in which the first current is one that flows through this
floating-regulator.
[0040] Further practices are those in which the first current
carries power and wherein diverting the first current includes
wasting this power.
[0041] In those cases in which the switches are in series along a
charge-transfer path, there exists practices of the invention in
which diverting the first current includes causing the first
current to avoid flowing along the charge-transfer path.
[0042] In another aspect, the invention features an apparatus
comprising a switching network. The switching network, when
connected to capacitors, forms a switched-capacitor network for
transforming a first voltage into a second voltage. The switching
network includes a plurality of switches, a corresponding plurality
of drivers, and a corresponding plurality of floating-regulators.
Among the switches is a first switch that connects to a floating
voltage. Among the drivers is a first driver to drive the first
switch using a drive voltage. A first floating-regulator from among
the regulators relies on voltage provided by an external voltage
source for causing the drive voltage to be at a fixed offset from
the floating voltage. This external voltage source is one that is
external to the switched-capacitor circuit. As such, it is not one
of the capacitors that forms the switched-capacitor network.
[0043] In some embodiments, the first floating-regulator includes a
first output and a second output. The first output is maintained at
a voltage that depends on the floating voltage; the second output
is maintained at a voltage that depends on the offset. In such
embodiments, first driver is connected between the first and second
outputs.
[0044] Also among the embodiments are those in which first
floating-regulator includes a first path that connects a voltage
provided by the external voltage source to the first driver and a
second path that connects a voltage corresponding to the offset to
the first driver. In some of these embodiments, the first path
includes a regulator switch and the second path includes a Zener
diode.
[0045] In some aspects, the invention features determining RDSON
values for each switch from a plurality of switches that, together
with a plurality of drivers that drives the switches and a
plurality of floating-regulators for controlling the drivers, forms
a switching network. The switching network, when connected to
capacitors, forms a switched-capacitor network for transforming a
first voltage into a second voltage. The switches are disposed in
series along a charge-transfer path. Among the switches are first
and second switches that connect to corresponding first and second
floating voltages. Determining RDSON values comprises causing a
known current to flow through the charge-transfer path and
concurrently determining RDSON values for each of the first and
second switches.
[0046] Some practices of the method include measuring a first
voltage across the first switch, measuring a second voltage across
the second switch, and based on the first and second voltages,
determining RDSON values for the first and second switches.
[0047] Also among the practices of the invention are those in which
causing a known current to flow through the charge-transfer path
includes inhibiting opportunities for the current to stray from the
charge-transfer path.
[0048] In another aspect, the invention features a switching
network that, when it is connected to capacitors, forms a
switched-capacitor network for transforming a first voltage into a
second voltage. The switching network includes a floating-regulator
that has two terminals. One terminal receives charge from one of
the capacitors and the other terminal receives charge from another
one of the capacitors. The floating-regulator uses charge from one
of the two capacitors to cause a driver to drive a switch from the
switched-capacitor network such that a driving voltage for the
switch is equal to a fixed offset from a floating voltage that is
present on one of the switch's terminals.
[0049] In another aspect, the invention includes a switching
network that, when connected to capacitors, forms a
switched-capacitor network for transforming a first voltage into a
second voltage. Such a switching network includes a switch, a
driver, and a floating-regulator. The floating-regulator includes a
first current path and a second current path. Current that proceeds
along the first current path drives the switch; current that
proceeds along the second path is diverted such that it fails to
provide power for driving any switch. The floating-regulator
includes a shunt for causing the current to proceed along one of
the first and second paths.
[0050] Some embodiments include a controller. In some of these
embodiments, the controller causes the shunt to select the second
path. In others, the controller determines whether a switch in the
switching network has a leakage current that complies with a design
specification. In these cases, the controller causes a bias current
in the switch, causes a diversion of this bias current so that the
bias current avoids interfering with measurement of the leakage
current, applies a voltage across the switch, obtains a measurement
of a current that flows in response to application of the voltage,
derives the leakage current from the measurement, determines that
the leakage current falls outside of the design specification, and
identifies the switching network as being defective. In still
others of these embodiments, the controller estimates a switch's
leakage current in part by causing a bias current that biases the
switch to be in a non-conducting state and diverting the bias
current so that it avoids interfering with measurement of the
leakage current.
[0051] In other embodiments, the second path leads to a node of
lowest electrical potential. This node has an electrical potential
that is lower than all other electric potentials in the switching
network.
[0052] Also among the embodiments are those that include a power
converter. Among these embodiments are those in which the switching
network and the capacitors define a switched-capacitor network that
is a constituent of the power converter. The power converter
includes a controller for controlling the switching network.
[0053] Also among the embodiments that include a power converter
are those in which the switching network and the capacitors define
a switched-capacitor network that is a constituent of the power
converter. During operation of the power converter, the shunt never
switches between the first and second paths.
[0054] In some embodiments, the switching network includes first
and second sets of switches. The switches from the first set are
arranged in series to define a first charge-transfer path that
extends from the switching network's input to its output.
Meanwhile, the switches from the second set of switches are
arranged in series to define a second charge-transfer path that
also extends from between the input and output. The first
charge-transfer path includes nodes at which switches from the
first set of switches connect to each other. These nodes connect to
a first set of capacitors. The second charge-transfer path likewise
includes nodes at which switches from the second set of switches
connect to each other. These nodes connect to a second set of
capacitors.
[0055] Also among the embodiments are those in which the switching
circuit is integrated into a monolithic substrate.
[0056] It is not necessary to have only one such substrate. In
other embodiments, the switching circuit is integrated into a first
monolithic substrate and the capacitors are integrated into a
second monolithic substrate that is interconnected with the first
monolithic substrate.
[0057] In other embodiments, the capacitors are not integrated into
a substrate. In these embodiments, the switching circuit is
integrated into the monolithic substrate and the capacitors are
connected to the switching circuit without having to be on a
monolithic substrate.
[0058] Embodiments further include those in which the switch is a
power FET and those in which the switch is one of several switches
that are distributed between first and second charge-transfer
paths. Both of these charge-transfer paths extend between an input
and an output of the switching network. In such embodiments, the
floating-regulator is cross-coupled between the first and second
charge-transfer paths.
[0059] Also among the embodiments are those in which the
floating-regulator includes a regulator switch to regulate flow of
charge into the driver and a Zener diode. The Zener diode causes
the flow of charge to be provided at a fixed offset from a floating
voltage.
[0060] In yet other embodiments, the floating-regulator includes a
regulator switch to regulate flow of charge into the driver and a
plurality of conducting paths for supplying currents to the
regulator switch. These conducting paths extend between different
sources of charge and the regulator switch.
[0061] Further embodiments include those in which the
floating-regulator includes a first diode that connects the
regulator switch to a first source of charge and a second diode
that connects a regulator switch to a second source of charge.
[0062] In still other embodiments, at least some of current that
proceeds along the first current path and that drives the switch is
provided to an additional floating-regulator to provide power for
driving an additional switch.
[0063] In another aspect, the invention features a method that
includes estimating a leakage current of a switch in a switching
network in an integrated circuit. The switching network, when
connected to capacitors, forms a switched-capacitor network for
transforming a first voltage into a second voltage. It includes a
driver and a floating-regulator. Estimating the leakage current
includes causing a first current, at least some of which will be
used to bias the switch, diverting the first current along a path
such that the first current avoids interference with a measurement
of a second current, obtaining a measurement of the second current,
and determining, the leakage current based at least in part based
on the measurement of the second current.
[0064] In some practices, the switch is one of a plurality of
switches that are in series along a charge-transfer path. In these
practices, diverting the first current includes causing the first
current to avoid flowing along the charge-transfer path.
[0065] Alternative practices of the invention include connecting a
voltage source across the switch. The voltage source is one that is
external to the switched-capacitor network. This means that it is
not one of the capacitors in the switched-capacitor network. The
second current is current that enters the switching network.
[0066] Yet other practices of include suppressing the ability of
the second current from flowing along a path other than a
charge-transfer path that leads to the switch.
[0067] In those cases in which the switch is a first switch that is
in series with a second switch along a charge-transfer path, there
are practices of the method that also include suppressing the
current that enters the switching network from flowing along the
charge-transfer path towards the second switch.
[0068] Additional practices include causing the first current to
flow through the floating-regulator, which controls driving of the
switch.
[0069] Also among the practices of the invention are those that
include wasting power by diverting the first current.
[0070] In another aspect, the invention features a switching
network that, when connected to capacitors, forms a
switched-capacitor network for transforming a first voltage into a
second voltage. The switching network includes a switch that
connects to a floating voltage, a driver configured to drive the
switch using a drive voltage, and a floating-regulator that relies
on a voltage provided by voltage source for causing the drive
voltage to be at a fixed offset from the floating voltage. This
voltage source is one that is external to the switched-capacitor
circuit. As such, it does not include a capacitor from the
switched-capacitor circuit.
[0071] Among the embodiments are those in which the
floating-regulator includes first and second outputs. The first
output is maintained at a voltage that depends on the floating
voltage and the second output is maintained at a voltage that
depends on the offset. The driver is connected between the first
and second outputs.
[0072] Also among the embodiments are those in which the
floating-regulator includes a first path that connects a voltage
provided by the external voltage source to the driver and a second
path that connects a voltage corresponding to the offset to the
driver.
[0073] In some embodiments, the floating-regulator includes a
regulator switch disposed along a first path and a Zener diode
disposed along a second path. The first path connects to a voltage
provided by a voltage source that is external voltage to the driver
and the second path connects to a voltage corresponding to the
offset to the driver.
[0074] Other embodiments include a controller configured to cause a
bias current to hold the switch in a conducting state and to divert
the bias current so that it avoids interfering with a known current
that is injected through the switch. The controller also causes
injection of the known current through the switch, measures a
voltage across the switch, and estimates an RDSON of the switch
based on the measurement of the voltage and the known current.
[0075] In still other embodiments, the switch is one of a plurality
of switches that are arranged in series to define a first
charge-transfer path that extends from an input of the switching
network to an output of the switching network. The first
charge-transfer path includes nodes at which switches from the
first set of switches connect. These nodes are configured to be
connected to the capacitors.
[0076] In another aspect, the invention includes estimating an
RDSON value for a switch that, along with a driver that drives the
switch and a floating-regulator for controlling the driver, is a
constituent of a switching network that when connected to
capacitors, forms a switched-capacitor network for transforming a
first voltage into a second voltage. The switch is disposed along a
charge-transfer path and connects to a floating voltage. Estimating
RDSON values includes causing a known current to flow through the
charge-transfer path and estimating an RDSON value for the
switch.
[0077] In some practices, the switch is one of a plurality of
switches in series along the charge-transfer path and the method
includes concurrently obtaining voltage measurements across each of
the switches and estimating the RDSON values based on the voltage
measurements.
[0078] Other practices include causing a known current to flow
through the charge-transfer path includes inhibiting opportunities
for the current to stray from the charge-transfer path.
BRIEF DESCRIPTION OF THE DRAWINGS
[0079] These and other features of the invention will be apparent
from the following description and the accompanying figures, in
which:
[0080] FIG. 1 shows a power converter that incorporates a
switched-capacitor network;
[0081] FIG. 2 shows details of the switched-capacitor network of
the power converter shown in FIG. 1;
[0082] FIG. 3 shows the switch circuit of FIG. 2 configured for
testing;
[0083] FIG. 4 shows the switched-capacitor network shown in FIG. 2
configured to test leakage current across a first pair of
switches;
[0084] FIG. 5 shows the switched-capacitor network shown in FIG. 2
configured to test leakage current across a second pair of
switches;
[0085] FIG. 6 shows the switched-capacitor network shown in FIG. 2
configured to test leakage current across a third pair of
switches;
[0086] FIG. 7 shows the switched-capacitor network shown in FIG. 2
configured to measure RDSON of three of the six switches;
[0087] FIG. 8 shows the switched-capacitor network shown in FIG. 2
configured to measure RDSON of the remaining three of the six
switches;
[0088] FIG. 9 shows steps for testing leakage current; and
[0089] FIG. 10 shows steps for testing RDSON.
DETAILED DESCRIPTION
[0090] FIG. 1 shows a power converter 2 for receiving an input
voltage VIN provided by a voltage source 3 and transforming it into
an output voltage VOUT that is made available at an output
capacitor 4 and a load 5 thereof. The power converter 2 includes a
controller 6 that controls a regulator 7 and a switched-capacitor
network 8. As used herein, the term a "switched-capacitor network"
includes a charge pump.
[0091] In the illustrated embodiment, the regulator 7 connects the
voltage source 3 to the switched-capacitor network 8. However, it
is also possible for the regulator 7 to connect the
switched-capacitor network 8 to the output capacitor 4. The power
converter 2 as shown can therefore be viewed as a specific
embodiment of a more general power converter that comprises a first
element connected to a second element, wherein when the first
element is a regulator, the second element is a switched-capacitor
network and wherein when the first element is a switched-capacitor
network, the second element is a regulator.
[0092] In some cases, the regulator 7 is implemented as a switched
inductor circuit, examples of which include a buck converter, a
boost converter, and a buck-boost converter. Such a
switched-inductor circuit can be operated with a duty cycle equal
to zero, thus transforming it into a magnetic filter. As such, in
some embodiments, the regulator 7 is simply a magnetic filter, such
as an inductor.
[0093] In other cases, it is possible to obtain adequate
performance with no regulator. Thus, embodiments of the power
converter 2 also include those with just the switched-capacitor
network 8.
[0094] Power converters of the type shown in FIG. 1 are described
in detail in U.S. Pat. Nos. 8,860,396, 8,743,553, 8,723,491,
8,503,203, 8,693,224, 8,724,353, 8,619,445, 9,203,299, 9,742,266,
9,041,459, U.S. Publication No. 2017/0085172, U.S. Pat. Nos.
9,887,622, 9,882,471, PCT Publication No. WO2017161368, PCT
Publication No. WO2017/091696, PCT Publication No. WO2017/143044,
PCT Publication No. WO2017/160821, PCT Publication No.
WO2017/156532, PCT Publication No. WO2017/196826, and U.S.
Publication No. 2017/0244318, the contents of which are all
incorporated herein by reference.
[0095] FIG. 2 shows the switched-capacitor network 8 in more
detail. The switched-capacitor network 8 includes a switching
circuit 10 that connects to first, second, third, and fourth
capacitors C1, C2, C3, C4. The switching circuit 10 is typically
integrated onto a monolithic substrate and connected, via pins, to
the capacitors C1-C4 that are either on another monolithic
substrate or that are provided as lumped elements.
[0096] The switching circuit 10 includes power FETS that implement
a first switch 12, a second switch 14, a third switch 16, a fourth
switch 18, a fifth switch 20, and a sixth switch 22. Each switch
12-22 has a current path that extends between its source and drain
and a gate that opens and closes the current path. In the
illustrated embodiment, the third switch 16 and the sixth switch 22
are implemented as PMOS transistors and the remaining switches are
implemented as NMOS transistors.
[0097] The first and second switches 12, 14 connect at a first node
N1A. The second and third switches 14, 16 connect at a second node
N2B. The fourth switch 18 connects to the fifth switch 20 at a
third node N2A. The fifth switch 20 connects to the sixth switch 22
at a fourth node N1B. Each of the nodes N1A, N1B, N2A, N2B connects
to a corresponding one of the capacitors C1-C4.
[0098] The first, second, and third switches 12, 14, 16 together
with the first and second nodes N1A, N2B define a first
charge-transfer path 24 that extends between an input VX and an
output node OUT_INT of the switched-capacitor network 8. The
fourth, fifth, and sixth switches 18, 20, 22 and the third node N2A
and the fourth node N1B define a second charge-transfer path 26
that likewise extends between the input VX and the output node
OUT_INT. A disconnect switch SD selectively connects and
disconnects the output node OUT_INT of the switching circuit 10 to
any other component. In FIG. 1, the component is an output
capacitor 4. However, in the testing procedure discussed below,
that component can be a voltage source or a voltmeter.
[0099] To control the first switch 12, it is necessary to control
the amount of charge on its gate. Accordingly, the first switch's
gate connects to a first driver 28.
[0100] The first driver 28 functions as a switch that drives charge
towards or away from the first switch's gate. To control the first
switch 12 quickly, the amount of charge on its gate must change
quickly. This requires that the first driver 28 provide a large
current into or out of the first switch's gate.
[0101] The first driver 28 switches between a first state and a
second state. In the first state, the first driver 28 receives
charge from a first floating-regulator 30 and provides it to the
first switch's gate. In the second state, the first
floating-regulator 30 removes charge from the first switch's gate
and disposes of it via the first charge-transfer path 24.
[0102] During normal operation, the charge that the first
floating-regulator 30 ultimately provides to the first driver 28
when it is in the first state comes from either the capacitor
connected to the first node N1A or the capacitor connected to the
third node N2A. This is advantageous because it is not always
possible to determine in advance which of these two capacitors will
be able to provide charge to meet the floating-regulator's
demand.
[0103] It is also possible to obtain charge from the voltage source
3 rather than from one of the two capacitors. This may occur during
start-up or testing.
[0104] Charge that comes from the capacitor connected to the first
node N1A reaches a first regulator switch 32 through a first diode
D1, the anode of which connects to the first charge-transfer path
24. Charge that the capacitor connected to the third node N2A
reaches the first regulator switch 32 through a second diode D2,
the anode of which connects to the second charge-transfer path 26.
Charge that comes from the voltage source 3 passes through the
third diode D3, the anode of which connects to the voltage source
3. The cathodes of the first, second, and third diodes D1, D2, D3
connect to the first regulator switch 32. Which of these three
sources of charge is ultimately used depends on which of the three
anodes is at the highest voltage.
[0105] The first and second diodes D1, D2 provide cross-coupling
between the first floating-regulator 30 and both the first and
second charge-transfer paths 24, 26. The first regulator switch 32
regulates the flow of this charge into the first driver 28. Charge
that ultimately reaches the first driver 28 during the first
driver's first state ultimately passes through the first regulator
switch 32.
[0106] As a result of the placement of the first and second diodes
D1, D2, the first floating-regulator 30 ultimately receives charge
during normal operation from either the capacitor connected to the
third node N2A or from the capacitor connected to the first node
N1A. This is advantageous since it is difficult to predict which of
these capacitors will be able to provide charge to supply the first
floating-regulator's demand for current.
[0107] In order to close the first switch 12, it is necessary to
create a conducting path between its drain and source terminals. To
do so, it is necessary to deposit enough charge on its gate to
raise its gate voltage past its source voltage by an amount that is
sufficient to form the conducting channel. Therefore, in addition
to depositing charge quickly onto the gate, the first driver 28
must also deposit the correct amount of charge. If the first driver
28 fails to deposit enough charge, the conducting channel will be
too small and hence excessively resistive. If the first driver 28
deposits too much charge, the first switch 12 can be damaged.
[0108] The correct amount of charge is one that raises the gate's
voltage past the source's voltage by a particular amount. A
difficulty that arises is that the source voltage floats. It is not
fixed. This means that the correct gate voltage cannot be fixed. It
must float with the source's voltage.
[0109] The first floating-regulator 30 achieves this by having a
reference point that is offset above the source's floating voltage.
In the illustrated embodiment, a Zener diode 34 creates this
reference point.
[0110] The Zener diode 34, a first resistor R1, and a second
resistor R2 together define a path that extends between the output
node OUT_INT and the first switch's source. The voltage drop along
this path is therefore the difference between the voltage at the
output node OUT_INT and the floating source voltage.
[0111] The anode of a fourth diode D6 connects to the voltage
source 3. Its cathode connects to the node at which the first and
second resistors R1, R2 also connect. This is useful to avoid a
potential short-circuit with the voltage source 3. The fourth diode
D6 and the third diode D3 cooperate during testing to provide power
to operate the first floating-regulator 30.
[0112] The Zener diode's anode connects to the first switch's
source and its cathode connects to the second resistor R2. As such,
the voltage at the Zener diode's cathode will always be a fixed
offset above the first switch's source voltage. The remainder of
the voltage drop, if any, between the source voltage and the output
node OUT_INT is borne by the first and second resistors R1, R2.
[0113] The Zener diode 34 cannot itself carry significant current.
However, it can be placed in parallel with a path that can carry a
significant current. If this is done in such a way that the voltage
drop across such a path crudely tracks the Zener diode, a voltage
with the correct offset above the floating source voltage can be
generated.
[0114] The Zener diode 34, in effect, functions as a voltage source
that maintains a voltage that exceeds the floating source voltage
by some fixed offset, the value of which depends on the VI
characteristic of the Zener diode 34. This voltage, when made
available to the gate of the first regulator switch 32, enables the
first regulator switch 32 to provide a voltage to the first driver
28, when it is switched into its first state, suitable for forming
a suitable conducting channel, thus closing the first switch
12.
[0115] A significant feature of the first floating-regulator is the
first-regulator shunt TM3. During normal operation of the
switched-capacitor network 8, the switching circuit 10 connects to
the capacitors C1, C2, C3, C4 and is actively engaged in converting
one voltage into another. In this case, the first-regulator shunt
TM3 is non-conducting. As a result, the current that biases the
first regulator switch 32 passes into the first charge-transfer
path 24.
[0116] However, shortly after being manufactured at a semiconductor
fabrication facility, it is desirable to test the switching circuit
10. In particular, it is desirable confirm that the leakage
currents and RDSON values for the various switches 12, 14, 16, 18,
20, 22 are within the design specifications. At some point during
this testing, the first-regulator shunt TM3 will be made to
conduct. As a result, the first-regulator shunt TM3 directs this
same biasing current to ground instead of to the first
charge-transfer path 24. This prevents the biasing current from
mixing with whatever current is already on the first
charge-transfer path 24. This also collapses the supply voltage to
the first driver 28.
[0117] This testing is expected to be the only time in the lifetime
of the switching circuit 10 that the first-regulator shunt TM3 will
be made to conduct. The first-regulator shunt TM3 is therefore only
used during testing.
[0118] The second switch 14 likewise has an associated second
driver 36 connected between a second floating-regulator 38 and the
first charge-transfer path 24 at the first node N1A. The second
floating-regulator 38 includes a second regulator switch 40 that
regulates the flow of this charge into the second driver 36 and a
Zener diode 42 connected to maintain a relatively constant gate
voltage across the second driver 36.
[0119] The second floating-regulator 38 receives current from
either the second node N2B or the fourth node N1B. Therefore, like
the first floating-regulator 30, the second floating-regulator 38
is cross-coupled between the first and second charge-transfer paths
24, 26. The structure and operation of the second
floating-regulator 38 is similar to that already discussed in
connection with the first floating-regulator 30.
[0120] Like the first floating-regulator 30, the second
floating-regulator 38 features a second-regulator shunt TM2. During
normal operation of the switched-capacitor network 8, the
second-regulator shunt TM2 remains non-conducting. As a result,
current for biasing the second regulator switch 40 flows into the
first charge-transfer path 24 and joins whatever current is already
there. At some time during testing, the second-regulator shunt TM2
will be made to conduct. This causes this current to pass to ground
instead of to the first charge-transfer path 24. Once testing is
complete, the second-regulator shunt TM2 is expected to never be
used again.
[0121] The output node OUT_INT supplies current directly to a third
driver 44 for charging the third switch's gate. When the third
driver 44 does not need current, this current goes to a third
floating-regulator 46. Additionally, when the third driver 44
removes current from the third switch's gate, that current also
goes to the third floating-regulator 46.
[0122] In either case, the third floating-regulator 46 recycles the
charge in this current by passing it to either the second node N2B
or to the fourth node N1B through which it ultimately reaches a
capacitor C2, C4. Therefore, like the first floating-regulator 30,
the third floating-regulator 38 is cross-coupled between the first
and second charge-transfer paths 24, 26.
[0123] The operation of the third floating-regulator 46 is slightly
different from that of the first and second regulators 30, 38
because the third switch 16 is implemented as a PMOS
transistor.
[0124] Like the first floating-regulator 30, the third
floating-regulator 46 includes a Zener diode 47 in series with a
resistor R3. The Zener diode again functions as a voltage source
that supplies a voltage equal to the third switch's source voltage
with a fixed offset. Since this Zener diode is in parallel with the
third driver 44, this voltage ensures placement of the correct
amount of charge on the third switch's gate to close the
switch.
[0125] Also, like the first floating-regulator 30, the third
floating-regulator 46 includes first and second diodes D4, D5.
However, instead of supplying charge from capacitors, these diodes
D4, D5 supply charge to capacitors. The first diode D4 supplies
charge to a capacitor connected to the fourth node N1B and the
second diode D5 supplies charge to a capacitor connected to the
second node N2B. The charge in both cases comes from the third
driver 44.
[0126] The third floating-regulator 46 features a third-regulator
shunt TM1. During normal operation of the switched-capacitor
network 8, the third-regulator shunt TM1 remains non-conducting. As
a result, current that enters the third floating-regulator 46
ultimately drains to either the second node N2B or to the fourth
node N1B and into one of the first and second transfer-paths 24,
26. At some time during testing, the third-regulator shunt TM1 will
be made to conduct, thus shunting current to ground instead of to
one of the first and second charge-transfer paths 24, 26.
[0127] The switching circuit 10 also has drivers and regulators for
the switches in the second transfer path 26. These are mirror
images of the corresponding strictures along the first
charge-transfer path 24. Accordingly, descriptions thereof are
omitted to avoid prolixity.
[0128] After having manufactured the switching circuit 10, it is
desirable to test one or more switches to confirm that those
switches meet design specifications for both leakage current and
RDSON. FIGS. 9 and 10 shows methods for doing so.
[0129] Part of manufacturing the switching circuit 10 is to test
the leakage current and RDSON of each switch 12, 14, 16, 18, 29,
22. This includes connecting the switching circuit 10 to a test
controller 48 as shown in FIG. 3.
[0130] Referring first to FIG. 4, to test the leakage current of
the first switch 12, the test controller 48 opens the first switch
12.
[0131] As can be seen in FIG. 4, the first floating-regulator 30
receives current from the output node OUT_INT. In normal operation,
this current would exit the first floating-regulator 30 and return
to a node of lower potential via the first charge-transfer path 24.
However, because the leakage current will also be flowing through
the first charge-transfer path 24, this extra current will pollute
the measurement of leakage current. To prevent pollution of the
measurement by this extra current, the test controller 48 closes
the first-regulator shunt TM3. This enables current leaving the
first floating-regulator 30 to bypass the first charge-transfer
path 24 on its way to a node of lower potential. As a result, the
only current that will flow through the first charge-transfer path
24 will be the leakage current that is to be measured.
[0132] The test controller 48 connects a current meter 85 between a
voltage source and the first node N1A. The voltage source then
applies a voltage at the drain of the first switch 12. Since the
source of the first switch 12 is grounded by that test controller
48, a test voltage develops across the first switch's drain and
source terminals. The test voltage is selected based upon the
voltage rating of the switch under test.
[0133] In response to this applied voltage, a leakage current flows
through the opened first switch 12. This leakage current draws
current through the current meter 85 towards the first node N1A. To
the extent this current all flows along the first charge-transfer
path 24 toward the first switch 12, it is equal to the leakage
current.
[0134] However, one cannot take it for granted that all current
passing through the current meter 85 will be leakage current. It is
apparent from inspecting the circuit's topology that when current
reaches the first node N1A, it will have three choices on where to
go next, only one of which is in the direction of the first switch
12.
[0135] To obtain an accurate measurement, it is important that,
upon reaching the first node N1A, all of this current be directed
toward the first switch 12. This is carried out by placing various
impediments to current flowing through any path but the desired
path.
[0136] The most significant alternative path is through the second
switch 14. Thus, the test controller 48 opens the second switch 14
to discourage current entering the first node N1A from going
through the second switch 14.
[0137] However, this is not enough. The second switch 14 is a
MOSFET. Therefore, has an inherent body diode. The orientation of
this body diode's cathode and anode creates the possibility that
current entering the first node N1A will flow through the body
diode instead of through the first switch.
[0138] To suppress this possibility, the test controller 48 also
applies a voltage at the drain of the second switch 14. This biases
the body diode and thus discourages entry of current that enters
the first node N1A, thereby encouraging all current entering the
first node N1A to pass through the first switch 12. As a result,
when the current meter 85 reports its measurement, that measurement
will indeed reflect the leakage current and not the sum of the
leakage current and some other current that ultimately went
elsewhere.
[0139] As shown, the voltage applied at the second node N2B is 26
volts and the voltage applied at the first node N1A is 20 volts,
thus leading to a six-volt bias across the second switch 14. For
those cases in which the switches are implemented as MOSFETs, it
would also be possible to apply 20 volts at the second node N2B in
which case the net applied voltage across the second switch 14
would be zero. This would still suppress current flow through the
second switch 14 because would have an intrinsic body diode that
would block current flow driven until the voltage driving that
current flow increases beyond the potential barrier imposed by that
body diode's space-charge layer.
[0140] The remaining alternative path is one that leads to the
first floating-regulator 30 and a fourth floating-regulator
connected to the fourth switch 18. However, it is apparent from
inspection of the topology that this path will already be blocked
because the first regulator switch 32 and a regulator switch in the
fourth floating-regulator will have been opened.
[0141] FIG. 4 also shows the test controller 48 testing the leakage
current of the fourth switch 18 at the same time. Such testing is
carried out using the same methods used in connection with testing
the leakage current of the first switch 12.
[0142] FIG. 5 shows the test controller 48 concurrently testing the
leakage currents at the second switch 14 and at the fifth switch
20. The same principles discussed in connection with the leakage
current measurement of the first and fourth switches 12, 18 in FIG.
4 are used in the measurement shown in FIG. 5.
[0143] In FIG. 5, the test controller 48 closes the
second-regulator shunt TM2 for the same reason it closed the
first-regulator shunt TM3. The test controller 48 applies 40 volts
to the drain voltage of the disconnect switch SD and 36 volts to
the drain of the second switch 14. As a result, the net voltage
across the source and drain terminals of the third switch 16 is
four volts minus the body diode drop of the disconnect switch SD.
This nevertheless discourages current flow through the body diode
of the third switch 16 because the body diode's own space-charge
layer will create a potential barrier to suppress such current
flow.
[0144] FIG. 6 shows the test controller 48 concurrently testing the
leakage currents at the third switch 16 and at the sixth switch 22.
The same principles discussed in connection with the leakage
current measurement of the first and fourth switches 12, 18 in FIG.
4 are used in the measurement shown in FIG. 6.
[0145] In addition to measuring leakage current, the test
controller 48 also measures RDSON for each switch 12-22.
[0146] Closing a switch 12-22 forms a conducting channel between
the source and drain. It is through this channel that charge flows.
However, this channel, being made from a semiconductor, has a
non-negligible resistance. This resistance is RDSON. If RDSON is
too high, energy will be lost through heating. As a result, it is
important to confirm that the RDSON value is consistent with design
specifications.
[0147] A useful way to measure RDSON is apply a suitable gate
voltage so as to create the conducting channel, to inject a known
current through this channel, and to then measure the voltage
between the source terminal and the drain terminal. The ratio of
this voltage measurement and the known current provides a way to
determine RDSON.
[0148] This method is particularly useful for the circuit shown in
FIG. 2 because each charge-transfer path 24, 26 passes through the
conducting channels of three switches 12-16, 18-22. This makes it
possible to inject one known current into the charge-transfer path
24, 26 and to then concurrently measure the voltages across three
switches 12-16, 18-22.
[0149] FIG. 7 shows the test controller 48 making an RDSON
measurement for the switches 12-16 on the first charge-transfer
path 24.
[0150] The test controller 48 closes all the switches 12-16 on the
first charge-transfer path 24 and opens all the switches 18-22 on
the second charge-transfer path 26. This ensures that the known
current flows through only the first charge-transfer path 24. A
bias voltage source VT connected to the output causes the source
voltages of all the switches 12-16 on the first charge-transfer
path 24 to be close a bias voltage VT. Because of this bias
voltage, the voltage provided to the first and second regulators
30, 38 to provide power to their respective gate drivers 28, 36 is
the bias voltage VT increased by the gate voltage VG required to
close the relevant switch 12, 14, 16.
[0151] The test controller 48 then connects first, second, and
third voltmeters 90, 92, 94 across the drain and source terminals
of the first, second, and third switches 12, 14, 16
respectively.
[0152] The test controller 48 then connects a test current source
IT so as to force a known current through the first charge-transfer
path 24. As a result of this current, the bias voltage VT measured
by the first voltmeter 90 will drop by a small amount
V.DELTA.4-V.DELTA.3 as a result of the RDSON of the first switch
12. The second and third voltmeters 92, 94 will likewise measure
small drops V.DELTA.3-V.DELTA.2 and V.DELTA.2-V.DELTA.1 caused by
the RDSON values of the second and third switches 14, 16.
[0153] It is important that the current that actually flows through
the first charge-transfer path 24 in fact be equal to the current
provided by the current source IT. This means that no current
should enter or leave the first charge-transfer path 24.
[0154] When measuring leakage current, the most significant source
of inaccuracy arose from current that was being used to bias the
floating-regulators. These currents would eventually find their way
into first charge-transfer path 24. Once there, these currents
would add to the leakage current that was already there. Since the
leakage current was small, even this stray current was enough to
taint the measurements of leakage current.
[0155] When measuring RDSON, the known current provided by the test
current source IT is quite large. As such, the entry of stray
currents used to bias the floating-regulators matters less.
However, it is still possible for some of this test current to
either stray away from the first charge-transfer path 24 or to fail
to enter the first charge-transfer path 24 in the first place. In
either case, the result is a poorer estimate of RDSON.
[0156] To ensure that all the test current at least enters the
first charge-transfer path 24, it is useful to open all the
switches that are on the second charge-transfer path 26, namely the
fourth switch 18, the fifth switch 20, and the sixth switch 22.
[0157] Once the test current is on the first charge-transfer path
24, it is still possible for it to prematurely leave the path by
entering one of the first and second floating-regulators 30, 38. To
avoid this, it is useful to reverse bias the first and second
diodes D1, D2 of each of the first and second floating-regulators
30, 38.
[0158] The process of reverse biasing includes applying, to the
anode of the third diode D3 of each of the first and second
floating-regulators 30, 38, a voltage that is in excess of the bias
voltage VT. In the illustrated example, this excess over the bias
voltage VT is selected to be the gate voltage VG. As a result, the
applied gate voltage VG is doing two things at once: it is biasing
a switch and also confining test current in the first
charge-transfer path 24 by suitably biasing the first and second
diodes D1, D2, thereby barring the entry of any test current into
the first and second floating-regulators 30, 38.
[0159] FIG. 8 shows the test controller 48 configured to measure
RDSON for the switches 18, 20, 22 on the second charge-transfer
path 26. The configuration is essentially the mirror image of that
described in connection with the switches 12, 14, 16 on the first
charge-transfer path 24.
[0160] In some cases, it is also useful to measure RDSON of the
disconnect switch SD as well as the switches within the switching
circuit 10. The test controller 48 carries this out by connecting a
fourth voltmeter 96 between a testing pin 98 and the output of the
switched-capacitor network 8 and then closing a testing switch 100
that couples the testing pin 98 to the output node OUT_INT of the
switching circuit 10. This testing switch 100 remains open during
normal operation and is only closed for this testing procedure.
Closing the testing switch 100 places the fourth voltmeter's
terminals on the drain and source of the disconnect switch SD.
[0161] The test controller 48 implements a method for testing
switches.
[0162] The method described and claimed herein results in an
improvement to a technological field, namely the field of testing
semiconductor devices. To the extent the set of all such methods
can be divided into abstract methods and non-abstract methods, the
methods as described herein are only the non-abstract methods. Any
descriptions of abstract methods have been specifically excluded
from this specification. In addition, the methods described herein
are only those that cannot be carried out by a human being using
only a writing implement and paper. Thus, none of the methods
described and claimed herein are purely mental steps.
[0163] FIG. 9 shows a non-abstract testing method 50 for estimating
the leakage current through a switch 12.
[0164] The process begins by opening the switch 12 whose leakage
current is to be estimated (step 52) and applying a known voltage
across it. In response to this voltage, a small leakage current
will flow through the switch 12 (step 54).
[0165] However, the process of controlling switches 12 itself
requires current. These currents pass through the various
floating-regulators and drivers that control the switches.
Collectively, these currents will be referred to herein as "biasing
current" to distinguish them from the "leakage current" that is of
interest. In normal operation, these biasing currents ultimately
make their way to the same path that goes through the switch
12.
[0166] Given the relative magnitudes of the biasing current and the
leakage current that is to be measured, it is important that this
current not be allowed to mix with the leakage current. Otherwise,
the estimate of leakage current will be degraded.
[0167] To maintain the purity of the leakage current, the method
includes the step of diverting the biasing current so that it does
not mix with the leakage current that is to be measured (step 56).
Typically, the biasing current is simply diverted to ground. One
way to do so is to close corresponding shunts TM1, TM2, TM3. In
either case, a result of this diversion, the biasing current will
not mix with the leakage current.
[0168] The process then continues with measuring the current
through the switch (step 58). One way to do this is to place a
current sensor along that current path to sense the amount of this
leakage current.
[0169] Because the current through the switch is now uncontaminated
by current from any other source, it provides a basis for
estimating the leakage current (step 60). The method then proceeds
with determining whether the estimate of leakage current is within
a design specification (step 62). If it is not, then the switching
network is rejected (step 64). Otherwise, the process determines if
there are more switches to test (step 66) and if so, proceeds to
test the next switch (step 52). Otherwise, the circuit is passed
(step 68).
[0170] FIG. 10 shows a process 70 for estimating the RDSON of
several switches 12, 14, 16 at the same time.
[0171] This process begins by closing the switches 12, 14, 16 that
are to be tested (step 72) and causing a known test current to
enter a first charge-transfer path 24 that passes through all of
the switches 12, 14, 16 (step 74). The process also includes taking
steps to prevent that test current from escaping the first
charge-transfer path 24 until it has passed through every switch
12, 14, 16 (step 76).
[0172] The testing current that enters the first charge-transfer
path 24 is prone to escaping by passing through floating-regulators
that are used to drive the switches. One way to suppress such
escape is to prevent entry of any portion of the testing current
into a floating-regulator. This can be carried out by applying a
voltage that reverse biases a diode that would otherwise admit
current into the floating-regulator. This applied voltage that
biases this diode can also be used in connection with driving the
switch.
[0173] The voltages across the switches 12, 14, 16 can then be
measured at the same time or at essentially the same time (step
78). Based on these measured voltages and the known current, it is
possible to estimate the RDSON of each switch 12, 14, 16 (step 80).
This is typically carried out by dividing the measured voltage by
the known value of current. The quotient that results from dividing
this measured voltage by the known current is an estimate of
RDSON.
[0174] The process proceeds with comparing the voltages to a design
specification (step 82). If any voltage is outside the design
specification, the circuit is rejected (step 84). If not, the
process determines if any more switches are to be tested. If there
are none, the circuit is marked as having completed the RDSON test
(step 86). Otherwise, the process continues with testing the next
switches.
[0175] As used herein, a "controller" refers to a tangible piece of
hardware that consumes electrical energy in the course of
performing work that includes moving electrical charge. The
controller is made of a combination of baryonic matter and leptons.
The controller generates waste heat and thus warms its environment.
The controller has mass and is not an intangible structure. Nor is
the controller software per se.
[0176] In some implementations, a tangible and non-transitory
computer-readable storage-medium includes a database representative
of one or more components of the power converter 2. Among these are
implementations in which the database includes data representative
of a switching circuit 10 that has been optimized to promote
low-loss operation of the switched-capacitor network 8.
[0177] As used herein, a computer-readable storage-medium includes
any non-transitory storage media accessible by a computer during
use to provide instructions and/or data to the computer. Examples
of computer-readable storage-media include storage media such as
magnetic disks, optical disks, and semiconductor memories. These
are non-abstract structures that are made of matter having
interacting baryons and leptons.
[0178] In particular embodiments, a database representative of the
system is a database or other data structure that is readable by a
program and used, directly or indirectly, to fabricate the hardware
comprising the system. The database is manifested in the real world
by rearrangements of certain attributes of matter such as charge
and direction of spin.
[0179] One example of such a database is a behavioral-level
description or register-transfer level (RTL) description of the
hardware functionality in a high-level design language (HDL) such
as Verilog or VHDL. The description may be read by a synthesis tool
that may synthesize the description to produce a netlist comprising
a list of gates from a synthesis library. The netlist comprises a
set of gates that also represent the functionality of the hardware
comprising the system. The netlist may then be placed and routed to
produce a data set describing geometric shapes to be applied to
masks. The masks may then be used in various semiconductor
fabrication steps to produce a semiconductor circuit or circuits
corresponding to the system. In other examples, alternatively, the
database may itself be the netlist, with or without the synthesis
library, or the data set.
[0180] Various features, aspects, and embodiments of
switched-capacitor power-converters have been described herein. The
features, aspects, and numerous embodiments described are
susceptible to combination with one another as well as to variation
and modification, as will be understood by those having ordinary
skill in the art. The present disclosure should, therefore, be
considered to encompass such combinations, variations, and
modifications.
[0181] Additionally, the terms and expressions that have been
employed herein are used as terms of description and not of
limitation. There is no intention, in the use of such terms and
expressions, of excluding any equivalents of the features shown and
described, or portions thereof. It is recognized that various
modifications are possible within the scope of the claims. Other
modifications, variations, and alternatives are also possible.
Accordingly, the claims are intended to cover all such
equivalents.
* * * * *